I169 trying nimble parsec

[RobertDober]

@bradhanks
So I was kind of shocked that I was too stupid to use NimbleParsec, therefore I tried again ;)

Seriously now, I would be interested in what you think about NimbleParsec. I am well aware that this still tiny code does not address the many challanges we would have in trying to integrate NimbleParsec into EarmarkParser

1. How to refactor for better readability

2. How to create helpers to not repeat things like " " |> string() |> times(min: 1) |> ignore()

3. How to feed enough lines into NimbleParsec, probably not really possible, thusly we need to join the rest of the lines, pass it into the HTML parser, and resplit and retokenize the rest, which would probably invalidate the "split the whole document into lines at first" approach (which I never really liked).

   We could probably already now just tokenize each line on demand, like creating a stream of lines by advancing to each /\n\r?/ when needed (would optimize inline code parsing too)

[RobertDober]

this is motivated by #169

[bradhanks]

I'm taking a look into nimble_parsec. My ambitions currently are to replace the Regex for attribute parsing. Different ways to improve parsing whole HTML blocks will probably seem less intimidating at the tail end of whatever we decide on this first pass.